Using Generalized Procrustes Analysis (GPA) for normalization of cDNA microarray data

BMC Bioinformatics

Table 2 Comparison among different normalization methods based on simulated normal microarray data with the SIMAGE method. The data sets include 1000 genes in 50 slides with 5, 10, 30% differentially expressed genes and the ratio of up-regulated to down-regulated genes is 1:1.

Method	5%⁽¹⁾		10%⁽¹⁾		30%⁽¹⁾
	ν ⁽²⁾	β ⁽²⁾	ν	β	ν	β
Raw	1.677	0.06525	1.678	0.08937	1.773	0.1615
Global	0.6751	0.06319	0.4611	0.08158	0.5158	0.1697
Loess	0.2089	0.05962	0.1863	0.07056	0.1891	0.1642
Scale	1.273	0.06859	1.145	0.09515	1.368	0.1921
Quantile	0.2546	0.06579	0.2085	0.07702	0.2167	0.1825
VSN	0.2331	0.05441	0.177	0.06695	0.2014	0.1573
GPA	0.08605	0.04569	0.09839	0.06639	0.1063	0.1328
Global+Scale	0.5449	0.06481	0.3928	0.08096	0.4448	0.1851
Global+Quantile	0.08674	0.04529	0.2085	0.07702	0.2167	0.1825
Global+GPA	0.2546	0.06579	0.09967	0.06784	0.107	0.13
Loess+Scale	0.1846	0.06	0.171	0.06968	0.1788	0.161
Loess+Quantile	0.2055	0.06124	0.1852	0.07132	0.1895	0.1638
Loess+GPA	0.1324	0.0536	0.1221	0.06078	0.1291	0.1468

ISSN: 1471-2105